Post on 09-Dec-2021
Frith - Neurocognitive basis of autism
The neurocognitive basis of autism
Uta Frith
The cognitive study of the underlying mental abnormalities in autism has advanced
rapidly, while the biological study of the underlying brain abnormalities and of
putative genetic mechanisms is lagging somewhat behind. However, the linking of
cognitive and biological studies has become a real possibility. Developmental cognitive
neuroscience has transformed our understanding of this enigmatic disorder, which was
once misguidedly thought to be caused by maternal rejection. The hypothesis of a
specific theory of mind deficit was a crucial step in this process. It explains the puzzle
of the characteristic social and communication impairments of autism and allows for
the fact that they can coexist with good general abilities. This hypothesis has been
widely accepted and a start has been made at pinpointing the brain basis of theory of
mind. The non-social impairments of autism have now become a major focus for
cognitive research. One theory proposes dysfunction in executive processes, in an
attempt to explain repetitive behaviour and inflexibility. Another theory proposes
weak information integration, in an attempt to explain narrow interests and special
talents. Autism research has thus stimulated ideas on important mind-brain systems
that may be dedicated to the development of social awareness, executive functions and
integrative processing.
A utism holds a peculiar fascination from which neuro-
scientists have not been immune. This fascination is re-
flected in the publication of three major field reviews in
the last two years’-‘. Why does a rare disorder excite such
interest? And what are the advances in research that
prompted the recent reviews?
Autism was first labelled by Kanner” and Aspergers in
the forties and has had a firm place in psychiatric textbooks
ever since. Many theories about the origin of autism have
been put forward and these have gradually become focused
on genes, brain and mind”. But in what way has the expla-
nation of autism progressed? Both cognitive and biological
studies have contributed to a better understanding.
However, it is the promise of linking together the cognitive
and biological aspects of autism in a principled way, that
makes the field so particularly exciting.
A rare disorder?
Autism, defined as a psychiatric disorder of early onset with
marked aloofness and obsessive desire for sameness, was
found to occur in four out of every 10 000 births’. These
inspired, if somewhat arbitrary, criteria have now been
broadened on the basis of better knowledge of the nature
and course of the disorder. Given these broader criteria,
recent studies show that autism affects between 10 and 30
individuals in every 10 000, that is, O.l-0.3%. Males pre-
dominate at a ratio of -2.5 :l, and this ratio increases with
higher levels of abilitya. Whether or not this reflects an in-
crease in the condition remains conjecture’. The popu-
lation estimates do nor include individuals with more sub-
tle forms of impairment. A Scandinavian study suggests
that the prevalence of this variant, now labelled Asperger
syndrome, might be as high as 0.5-0.7% (Ref. 10).
Individual differences in the clinical picture
The chief diagnostic criteria for autism, as set out in ICD-10
(Ref. 11) and DSM-IV (Ref. 12), are abnormalities of social
interaction, impairments in verbal and non-verbal commu-
nication and a restricted repertoire of interests and activ-
ities, all being present from early childhood. The diagnostic
criteria for Asperger syndrome overlap, but stipulate that
there should be no significant developmental delay in cog-
nitive and language functions.
Three-quarters of the individuals affected by autism
suffer from mild to severe mental retardation. Measured IQ,
at all levels, shows marked peaks and troughs on different
subtests”,“. The presence of savant skills is as high as 10%
in the autistic population, and conversely, almost all savants
are diagnosed as suffering from autism”. The savant is de-
fined as an individual with outstanding skill in one of a
small range of areas of interest, including calendar calculation
and high musical or artistic competence, while his other
intellectual abilities can be at a very low level. Examples of
drawings by a savant artist are shown in Fig. 1 (Ref. 15).
Copyright 0 1997, Elsevler Science Ltd. All rights reserved. 1364.6613/97/$17.00 PII: 51364.6613(97)01010-3:
Trends in Cognitive Sciences - Vol. 1. No 2, May 1997
Frith - Neurocognitive basis of autism
Fig. 1 Exceptional drawing ability is shown in these exam-
ples of E.C., a ‘savant’ with autism, who has been extensively
studied by Laurent Mottron and Sylvie Belleville’s. The upper
drawing is spontaneous; the lower drawing was made when
asked to ‘turn them around in your head and draw them’. To
produce this drawing E.C. started with the lower part of
the circle. He only stood up to examine the back of the
objects for the upper part of the circle. In the spontaneous
drawing, some objects are left incomplete, stopping at the
edge of the sheet. E.C. frequently starts a drawing with an
apparently arbitrary detail of an object and may leave objects
or persons half-completed or with an essential part missing.
He never corrects a line once it is drawn. Reproduced with per-
mission from Ref. 15.
The extent of individual variability in signs and symp-
toms, and intellectual abilities, has only recently become ap-
preciated fully’i. The interaction between specific disabil-
ities and general compensatory processes might account for
some of the variability, but has hardly been investigated.
Thus, some individuals, notably those with more severe
forms of mental retardation, never develop speech or other
means of communication, while others acquire fluent lan-
guage”. However, even well compensated individuals will
show some difficulty with the to-and-fro of everyday chat.
This will be noticeable in the prosody, style and content of
speech. Obsessive and repetitive tendencies, as well as nar-
row interests, also vary and are prevalent in individuals of
both low and high general ability. They often impose a
heavy burden; long neglected, these features are now receiv-
ing renewed research interest’3,‘7.
The developmental course
The clinical picture of autism differs between individuals
and changes markedly over the course of development, with
a tendency towards improved adaptation. Nevertheless, the
underlying cognitive disorder (and by implication, brain
abnormality) appears to remain throughout life. It is during
early childhood that all the characteristic signs of autism are
most strongly apparent, in particular, poverty of imaginative
play and lack of social-emotional rapport with other chil-
dren. A simple screening instrument applied at 18 months
of age has been remarkably successful at predicting autism.
Identification of autism at this age is based on a failure to show
ordinary pointing for the purpose of shared attention, an ab-
sence of gaze monitoring and an absence of pretend play’y.
Autistic individuals with near normal or higher intellec-
tual ability, while not exempt from behavioural problems,
can show a high degree of improvement and adaptation,
and this may make their diagnosis controversial. One par-
ticular challenge is to explain why the prospects of leading
an independent normal life are poor, even in individuals
whose overt impairments are subtle. These individuals still
show problems with peer interaction and are extremely
unlikely to have intimate friends. They also tend to have
difficulties in planning their daily life and, perhaps as a
secondary consequence of their condition, may suffer from
high levels of anxiety or depression. An important question
concerns the factors contributing to compensation and
their costs to the individual. For instance, ‘passing for nor-
mal’ in social interactions may require continuous effort,
perhaps because non-intuitive and non-automatic infor-
mation processing strategies are required.
It is less difficult to explain the daily living problems of
autistic individuals who also suffer from mental retardation.
The most severe cases remain socially unaware, lack both
verbal and non-verbal communication, and often have chal-
lenging behaviour problems, including self-injury. The
management of these problems is largely based on behav-
iour modification techniques. Cognitive models have so far
contributed little to this area. Since cooperation in psycho-
logical experiments is hard to gain from this subgroup, rel-
evant research is rare.
Life span studies of autism covering middle and old age
are as yet lacking. Similarities between low-functioning
autistic individuals in middle age with negative symptom
schizophrenics have been pointed out. This observation
may yield clues to dysfunction at the neural level, even if
onset and course of the disorders are different”.
The autistic spectrum
Three primary features: characteristic impairments in social-
ization, communication and imagination were identified as
highly correlated in a geographically defined population
study’” of autism. They have been shown to emerge early
and persist in development even though their precise mani-
festation changes (Fig. 2). F or example, aloofness may give
way to social pestering. The triad of impairments appears
to be a common denominator throughout a spectrum of
autistic disorders.
The broader phenotype
Twin studies too have resulted in the recognition of a
broader autistic phenotype that confirms and extends the
notion of the autistic spectrum. The strong heritability of
autism is beyond doubt. Given a narrow definition of autism,
if one member of a pair of monozygotic twins has the disor-
der, then in over a third of cases the other twin also has it. In
contrast, such concordance is hardly ever found in dizygotic
twin pairs. When a wider definition of autism is used, allow-
ing for more subtle impairments, then the concordance rate
doubles, with 72% for monozygotic vs 10% for dizygotic
Trends in Cognitive Sciences - Vol. 1. NO. 2. May 1997
pairs*‘. Clearly, the clinically based category of autism cur-
rently used is rather strict compared to the phenotypically
derived category. This suggests that there may be hidden
cases. There could be as yet unknown factors that protect
some individuals from being affected by the clinically sig-
nificant disorder, despite a genetic predisposition.
Asperger syndrome
One of the most important changes in autism research and
practice in the last five years has been the growing recog-
nition of a subgroup of individuals with high verbal ability
and the potential for good social adaptation. Asperger syn-
drome is still a controversial concept and has uncertain
boundaries with other clinical categories, such as high-
functioning autism, atypical autism and schizoid personal-
ityayZL,ai. Typically, the diagnosis ofAsperger syndrome is made
in later childhood, or even adulthood, even though abnor-
malities have been present from early in life. Sometimes, a
late diagnosis is made where the initial diagnosis was classi-
cal autism, and where language and social communication has
improved beyond expectation. Sometimes, a late diagnosis
is made where the initial assumption was ‘eccentricity’, tol-
erated within the family, but not outside. The Asperger
variant of autism is often called ‘mild’. This obscures the
fact that Asperger syndrome sufferers have distinct handicaps
in social and non-social spheres of everyday life, ranging
from mild to severe. Thus, an individual who is also very
able intellectually, can become a respected academic whose
communication problems, obsessive tendencies and restricted
interests may be overlooked; however, a less fortunate and
less able individual may be shunned by society and be a
danger to himself and others.
Biological studies
Behavioural genetic data suggest a small but significant re-
currence rate in families and favour a multiplicative, multi-
locus model of inheritance, perhaps involving only a small
number of genes’. However, no precise findings are as yet
available. Other disorders of known biological origin, such
as fragile-X syndrome, phenylketonuria and tuberous scle-
rosis, can lead to the clinical picture of autism, but this is
usually linked to severe mental retardation”.
Neurological and neuropsychological signs, such as
epilepsy, motor abnormalities and abnormal profiles on
psychological tests have always been noted. Recent studies
have highlighted the resemblance of the profiles to those of
patients with frontal lobe and right hemisphere lesion?~“‘.
Post-mortem brain studies have revealed a number of ab-
normalities in cell structure in the hippocampus, amygdala
and cerebellum. The findings suggest a curtailment of neur-
onal development at, or before, 30 weeks of gestation”. In
viuo brain imaging studies have also shown a number of ab-
normal patterns in brain structure and function in these and
additional regions. None of these studies are, however, spe-
cific or consistent enough to build up a coherent theoretical
picture of the origin and nature of the brain abnormality.
The theory of mind deficit hypothesis
Cognitive explanations of the core features of autism provide
a vital interface between brain and behaviour. The proposal
Frith - Neurocognltlve basis of autism
Fig. 2 Photograph of a young autistic child. Children with
autism are often fascinated by unusual objects. While the social
and communication impairments of autism have received a
great deal of research interest, the non-social features are now
beginning to be understood as well. This photograph is repro-
duced by kind permission of Sue’s family.
of a specific neurologically based problem in understanding
minds was a significant step in this endeavou?‘. The hy-
pothesis that autistic children lack the intuitive understand-
ing that people have mental states, was originally tested with
the Sally-Ann false-belief paradigm”‘. This hypothesis has
been confirmed in a number of studies, and the idea of a
theory of mind deficit in autism has been widely accepted”.
Individuals with autism fail to appreciate the role of mental
states in the explanation and prediction of everyday behav-
iour. Hence they lack a full understanding of deception,
and of those emotions which depend on monitoring other
people’s attitudes, such as pride.
Individuals with Asperger syndrome stand out from
other autism sufferers because they appear to possess an
understanding of mental states, although they achieve suc-
cess on theory of mind tasks at a much later age than normal
and, to do this, they require higher levels of verbal knowl-
edge than normally developing children”. This could be be-
cause they show extremely delayed development of the pre-
requisite neurocognitive mechanism or because they have
been able to work out an alternative non-intuitive theory of
mind. The relationship between better mental state under-
standing and better verbal ability in Asperger syndrome is
consistent with the suggestion that language acquisition
itself is facilitated by the ability to attribute intention to
other speakers’“. A recent PET scan study of story compre-
hension, revealed that an area in the left medial prefrontal
cortex, shown in Fig. 3, was active when normal volunteers
had to infer the mental states of the story characters”. This
Trends in Cognitive Sciences - Vol. 1, No. 2, May 1997
Frith - Neurocognitlve basis of autism
Fig. 3 A PET scan study by Fletcher eta/.33revealed that when normal volunteers read a story which critically involves ‘theory
of mind’, the highlighted area in the left medial prefrontal
cortex (Brodmann area 8/9) is activated.
remarkably circumscribed area was not active in individuals
with Asperger syndrome when studied under the same con
ditions”. Further studies are needed to elaborate these re-
sults; however, it is striking that another study using a dif-
ferent mental state inference task also found increased
activation in this area’?. as did a previous study on story
comprehension’“. Such findings may provide an exciting
lead to the brain-mind basis of self-awareness.
The executive dysfunction hypothesis
A specific deficit in theory of mind is unable to account for
the whole of the clinical picture of autism. This is particu-
larly true for those features of autism that are encompassed
by an obsessive desire for sameness. These features are being
successfully tackled by a theory which proposes a deficit in
executive functions. These functions include planning.
working memory, impulse control, shifting set and the
initiation and monitoring of acrioni7. All are thought to de-
pend on systems which involve prefrontal activity in the
brain and are typically impaired in a range of disorders that
are likely to involve defects in the frontal lobe, including at-
tention deficit disorder, obsessive compulsive disorder”,
Tourette syndrome, phenylketonuria and schizophrenia.
Poor performance on many executive function tasks has
been documented in autism”~“; for instance, individuals
with autism and Asperger syndrome often fail to inhibit
overlearned responses when these are no longer appropriate.
Poor performance on these tasks is directly related to stereo-
typed and rigid behaviour in everyday life as shown in
highly repetitive thought and action. Problems are also ob-
served in the initiation and monitoring of action. The
underlying brain abnormality is likely to involve a distrib-
uted system with strong neural connections between basal
ganglia, striatum and prefrontal cortex, such as the dopa-
mine system’O. This would fit in with the tentative brain
imaging results on theory of mind which highlighted the
medial prefrontal cortex. However, the relationship between
deficits in executive funcrions and theory of mind has yet to
be investigated across frontal lobe disorders.
The weak central coherence hypothesis
A cognitive theory which proposes weak information inte-
gration/central coherence 6’1,4’ addresses the islets of ability
and exceptional talents which are present in a significant
proportion of individuals with autism and Asperger syn-
drome. The theory links these assets to the striking phe-
nomenon of restricted interest. The attention of the autistic
individual is often captured by fragments or surface fea-
tures of objects and sensations that are usually of little inter-
est to normal people. In one particular case, the special
interest was the colour of the doors of juvenile magistrates
courts - not those of adult courtP! Performance peaks
in tests such as ‘block designs’ and ‘embedded figures’, as
well as outstanding feats of rote memory or detailed draw-
ing of objects’s, appear to demonstrate a preference for
segmental over holistic information processing. Evidence
that people with autism process information in an un-
usually piecemeal fashion, at both the behavioural and
physiological level, was provided by the brain imaging study
of Happe et al?” Here the autistic participants showed less
difference in performance, as well as in brain activity, for
processing meaningful stories versus meaningless text
(jumbled sentences) compared with normal controls. This
is consistent with rhe findings of a study where they failed
to integrate information so as to derive contextually relevant
meaning. For instance, when reading aloud: ‘the dog was
on a long lead’, they pronounced the word lead as ‘led”‘.
A study in the visual domain that may relate to lack of
integration showed that individuals with autism are less
susceptible to certain visual illusions: this may be because
they are less affected by the context surrounding the lines
to be judged-“.
The relationships between central coherence, executive
functions and theory of mind are as yet unclear. The brain
basis ofcentral coherence is an open question. One particularly
intriguing suggestion follows from the finding that the brain
is organized to sense the synchronicity of cells firing in dif-
ferent locations4”. This capacity is thought to provide the
necessary cues for the binding of information. Thus, at the
level of brain physiology, impairment in the sensing of syn-
chronicity could lead to lack of coherence in perceptual pro-
cessing. At the cognitive level this might be reflected as a
lack of coherent perceptual experience.
.,....,...,..__.........,...,.............,...................................................... . . . . .
References 1 Ciaranello, A.L. and C~aranello, R.D. (1995) The neurobiology of
infantile autism Annu. Rev. Neurosci. 18, 101-128
2 Bailey, A.J., Phillips, W. and Rutter, M. (1996) Autism: integrating
clinical, genetic, neuropsychological. and neurobiological perspectives
J. Chrld Psycho/. Psychiatry 37,89-l 26
3 Happe, F and Frith. U. (1996) The neuropsychology of autism Brain
Trends in Cognitive Sciences - Vol. 1, No. 2. May 1997
119, 1377-1400
4 Kanner, L. (1943) Autistic disturbances of affective contact New. Child
2, 217-250
5 Asperger, H. (1944) ‘Autistic psychopathy’ in childhood. Translated and
annotated by U. Frith, in Frith, U.. ed. (1991) Autism and Asperger
Syndrome, pp. 37-92, Cambridge University Press
6 Frith, U. (1989) Autism: Explaining the Enigma, Blackwell
7 Lotter, V. (1966) Epidemiology of autirtlccondltions in young children:
I. Prevalence Social Psychiatry 1, 124-137
8 Glllberg, C. and Coleman, M. (1992) The Biology of the Autistic
Syndromes (2nd edn), Mac Keith Press
9 Wing, L. (1996) Autistic spectrum disorders: no evidence for or against
an increase in prevalence Br. Med. J. 312, 327-328
10 Ehlers, 5. and Gillberg. C. (1993) The epidemiology of Asperger
syndrome. A total population study J. Child Psycho/. Psychiatry 34,
1327-1350
11 World Health Organization (1992) The /CD-IO Clamfication of Mental
and Elehavioural Disorders, World Health Orgamzation
12 American Psychiatric Association (1994) Diagnostic and Statistical
Manual of Mental Disorders (4th edn) (DSM-IV). American Psychiatric
Association
13 Happb, F. (1994) Autism: an Introduction to Psychological Theory, UCL
Press
14 Rimland, 6. and Fein. D. (1988) in The Exceptional Brain (Obler, L.K.
and Fein, D., eds). pp. 474-492, Guilford Press
15 Mottron, L. and Belleville. 5. (1993) A study of perceptual analysis in a
high-level autistic subject with exceptional graphic abilities Brain
Cogn. 23, 279-309
16 Frith, U. and HappC, F. (1994) Language and communication in
the autistic disorders Philos. Trans. R. Sot. London Ser. B 346,
97-104
17 Russell, J., ed. Autism as an Executive Disorder, Oxford University Press
(in press)
18 Baron-Cohen, 5 et al. (1996) Psychologacal markers in the detection
of autism in infancy in a large population Br. J. Psychiatry 168.
158-163
19 Frith, U. and Frith, C.D. (1991) Elective affimties in schizophrenia and
childhood autism, in Sock/ Psychiatry: Theory, Methodology and
Prach’ce (Bebbington, P., ed.), pp. 65-88, Transactions
20 Wing, L. and Gould, J. (1979) Severe impairments of social interaction
and associated abnormalities in children: epidemiology and
classification J. Autism Dev. Disord. 9, 1 l-29
21 LeCouteur. A. et al. (1996) A broader phenotype of autism, the cl~mcal
spectrum in twns J. Child Psycho/. Psychiatry 37, 785-801
22 Frith, U., ed. (1991) Aut!sm and Asperger Syndrome, Cambridge
University Press
23 Wolff, 5. (1995) Loners: The Lifepath of Unusual Children, Routledge
24 Smalley, 5.. Asamow, R. and Spence, M. (1988) Autism and genetics.
a decade of research Arch. Gen. Psychiatry 45, 953-961
25 Rumsey, J.M. (1992) in High-functioning Individuals with Autism
(Schopler, E. and Mesibov, G.B., eds). pp. 41-64, Plenum Press
26 Klin, A. et al. (1995) Validity and neuropsychological characterisation
of Asperger Syndrome: convergence with nonverbal learning
disabilities syndrome 1. Child Psycho/. Psychiatry 36, 1127-l 140
27 Ozonoff, 5. and Miller, J.N. (1996) An exploration of right-hemisphere
contributions to the pragmatic impairments of autism Brain Lang. 52,
41 l-434
28 Bauman, M. and Kemper. T. . edr (1994) The Neurobiology of Autism,
Johns Hopkins Press
29 Frith. U., Morton, J. and Leslie, A. (1991) The cognitive basis of a
biological disorder Trends Neurorci. 14. 433438
30 Baron-Cohen, 5.. Leslie, A. and Frith, U. (1985) Does the autistic child
have a ‘theory of mmd’? Cognition 21, 37-46
31 Baron-Cohen, 5.. Tager-Flusberg, H. and Cohen, D.J.. eds (1993)
Understanding Other Minds: Perspectives from Autism, Oxford
University Press
32 Happe, F. (1995) The role of age and verbal ability in the theory
of mind task performance of subjects with autism Child Dev. 66,
843-855
33 Fletcher, P.C. et al. (1995) Other minds m the brain: a functional
imaging study of ‘theory of mind’ in story comprehension
Outstanding questions
l The unravelling of the biological causes of autism is eagerly awaited.
Such work should clarify diagnosis through direct biological markers. It might lead also to new approaches to the treatment and/or prevention of autism. Important bridges between neurobiology and cognitive
processes, such as those underpinning ‘theory of mind’ and executive
functions, are currently being built by a number of different investigators. In this sense, the study of autism serves a much wider aim
than that of understanding the basis of the disorder. l The question of individual variability and compensation is notoriously
difficult. It will be fascinating to compare high functioning autism,
Asperger syndrome and normal eccentricity. Are there sharp dividing
lines? Are there also marked individual differences in the normal population in the critical abilities underlying the social and non-social impairments in autism? The systematic study of the most severely impaired individuals is overdue. It will be important to investigate how
well autistic and non-autistic individuals can be distinguished on specific
cognitive tests at moderate and severe levels of mental retardation. The study of the course of the disorder over time should eventually lead to
the identification of useful predictors of the quality of social adaptation and communication.
l The juxtaposition of impaired and intact abilities in autism appears to
support a modular model of the mind where specific deficits can be
confined to particular modules. The neurological system underlying theory of mind is close to being identified in functional brain imaging studies. A similar approach should also elucidate the non-social features
of autism. However, the effects of poor executive functions, weak
central coherence and general mental retardation are pervasive rather than modular. How do specific mental components and general
cognitive system5 interact? l Autism research has stimulated philosophical debate on self-
consciousness, since it provides an example of the failure to reflect on mental states. The further experimental study of this topic should
intensify with the development of ideas on subcomponents and precursors of theory of mind in other species. These ideas, in turn, might
stimulate neuropathological, neurochemical and neurogenetic research on the development of conscious experience. This provides a thrilling
prospect for a long-term research programme of far reaching
importance.
Cognition 57, 109-128
34 Happe, F. et al. (1996) ‘Theory of mind’ in the brain Evidence from a
PET scan study of Asperger’s syndrome NeuroReport
35 Gael. V. eta/. (1995) Modeling other minds NeuroRepori 6,1741-1746
36 Mazoyer, B.M. et al. (1993) The cortical representation of speech
J. Cog”. Neurosci. 5, 467479
37 Shallice, T . (1988) From Neuropsychology to Mental Structure,
Cambridge University Press
38 Pennington, B.F. and Ozonoff, 5. (1996) Executive functions and
developmental psychopathology J. Child Psycho/. Psychiatry 37,
51-87
39 Hughes, C., Russell, J. and Robbins, T.W. (1994) Evidence for executive
dysfunction in autism Neuropsychologia 32, 477492
40 Damasio, A.R. and Maurer, R.G (1978) A neurological model for
childhood autism Arch. Neural. 35, 777-786
41 Frith, U. and Happe, F. (1994) Autism: beyond ‘theory of mind’
Cognition 50, 115-l 32
42 Happe, F. (1997) Central coherence and theory of mind in autism:
reading homographs in context Br. J. Dev. Psycho/. 15. 1-12
43 Happb. F. (1996) Studying weak central coherence at low levels:
children with autism do not succumb to visual Illusions. A research
note J. Child Psycho/. Psychiatry 37, 873-877
44 Singer. W. (1993) Synchronization of cortical activity and Its putative
role in information processing Annu. Rev. Physiol 55, 349-374
Trends in Cognitive Sciences - Vol. 1, No. 2, May 1997